Tree structure and carbon cycling in a tropical rainforest under long-term drought

Abstract

Tropical forests are highly productive and biodiverse, exchanging more carbon with the atmosphere than any other terrestrial ecosystem and representing the apex of taxonomic and structural diversity on land. However, their high productivity is sensitive to climate, which for many tropical regions, and especially the Amazon, is predicted to become more extreme this century, with stronger and more frequent drought events likely to occur. Extreme drought has been shown to alter net carbon gain in Amazon forests, causing a decline or sometimes a reversal in the natural carbon sink. The long-term effects of increasing drought exposure are poorly understood, but are thought likely to substantially weaken or reverse the future Amazon carbon sink, with impacts at the scale of the global carbon cycle. Our understanding of the response by any forest to drought is hampered by observational opportunity. One way to resolve this is to manipulate water availability experimentally at large scale. This thesis uses one such 'ecosystem-scale' (1 ha) rainfall manipulation experiment in Amazon rainforest, combining new structural, ecological and ecophysiological measurements on a forest that has experienced long-term experimental drought, over twenty years. The new structural measurements comprise terrestrial laser scanning (TLS) data that are analysed to quantify detailed tree and forest structural metrics. The physiological measurements focus on using these novel structural datasets to estimate and scale woody tissue carbon dioxide effluxes from twig to forest. Finally, biomass and growth are analysed over the twenty-year experimental drought period, using the new structural data to refine the quantification and understanding of change in biomass and growth during the period of experimentally-imposed drought. Tree structural plasticity in response to drought may mitigate the sensitivity of tropical rainforests to climate change. Here, I find that tropical rainforest woody and leafy structural traits altered substantially under drought in a manner that may influence individual tree and whole forest hydraulic conductance, resource acquisition efficiency and heat transfer. The tree structural plasticity found in droughted trees had implications for tree metabolic scaling, resulting in lower whole tree wood CO2 efflux for droughted trees in comparison to equivalent sized Control trees. This alteration in metabolic scaling occurred on droughted trees despite minimal change in wood CO2 efflux rates per unit tree surface area. At the stand-scale, altered tree structure and high rates of tree mortality led to over 50% reduction in stand wood CO2 efflux. Predicted drought scenarios in the tropics has led to substantial interest in climatic thresholds that can lead to ecosystem tipping points. The high mortality rate and accompanying loss of aboveground biomass during this drought experiment has provided the opportunity to test potential ecosystem collapse under extreme drought. By analysing biomass dynamics, validated using TLS, and tree woody growth over 20 years of experimental drought, I find evidence that the ecosystem has reached a new stable state, with reduced rate of aboveground biomass loss due to mortality, and with stabilized growth rates. There is large uncertainty about how climate change will influence carbon cycling in the tropics. I contribute to the understanding of the woody component of the carbon cycle in tropical rainforests under drought stress, using novel TLS techniques to link structure with function on a tree and forest-scale. Show more